Episode 22: The AI Trust Bucket
If someone produces better work with AI, how do you know they're getting better at their job? Justin and Kellan talk about trust, accountability, and what it takes to help people use AI well. Along the way: a confident recommendation that needed another question, a camp agent named George, and knowing when ordinary process improvement is enough.
The Drip:
- The OpenAI research discussed this week: quality output, reasoning, and the question of whether people are building expertise.
- The Anthropic coding study and what people could explain after completing the task.
Inside The Bottle:
- A full trust bucket still needs boundaries.
- Why a promising demo needs to survive a typical month's work.
- Getting the leadership team aligned before choosing another tool.
- Meeting people where they are: three useful jobs for a camp agent.
- Keeping learning and accountability in the work, including the questions a manager will ask.
New episodes every Wednesday. Subscribe to The Weekly Fizz newsletter at http://tinybottleai.com/the-fizz.
